Levelling my lathe - a build log

Ah, but, I didn't finish there!

I've made some levelling feet for the base and set up the lathe and attempted to level the base.

Is the base level?

Not really!

It's maybe a bit hard to see, but the centre is pretty level, but both left and right ends slope upwards to the right.

I must say it was a pleasure to break out the engineers level as I've never used it since asking for it for a birthday present about 10 years back! I was sure I would need it...

If I've understood this level correctly each graduation represents 0.02mm / m. In which case the drift upwards is over 20 microns across the width of the (100mm) level. That's taking a guess at how for off centre the bubble is (10 graduations..). This does seem to be maxed out so it could be anything.

When on the mill bed, there was no rocking of the base, but it was definitely spinning around the centre not near the ends, so I expect it is mildly out - Of course as well as straightening up the lathe, this will have twisted the base a little.

I've had a bit of a go at levelling just one end and also looking at the levels on the lathe bed, but it was doing my head in so I gave up for the day.

Any advice will be much appreciated!

Iain

All that you need is the two planes in line with the Y axis to be in alignment. It can slope slightly to the headstock or to the tail stock with no issues.

You can wind the carriage down to the headstock end, and with the level placed for and aft on cross slide, adjust the head stock to level. Wind down to the tail stock end, and adjust to get that end level. A little winding back and forth and you will have it. That will get the bed running true and fairly close. Check the headstock, to see if the run of the headstock aligns to the run of the bed/carriage movement . Indicate a bar to run fairly true at the chuck end and at the end of it's length. A low pressure dti is a big help in this . Most plunger clocks have too much pressure and can give a false reading. I prefer low pressure finger indicators for this type of work. Check for parallel at the top of the bar and from one side. Your test bar wants to be no more than 10 diameters long.

After that it is a matter of setting up the tailstock to be true to the chuck and check it at a few different length positions along the lathe.. start close to the chuck , and adjust the centre to match to cut the same at each end on something that is short like a piece of 10-13mm ali bar out about 30mm. Then check with a more substantial bar over the full range of the saddle /tailstock travel. If this is out, you may need to adjust the tailstock feet a very small amount to get it in alignment. Then test the centre area of travel and then retest the front section again. When all 3 will produce the same result of a parallel cut, it's as good as it gets.

Most lathes have the tailstock up to0.05mm or so above centre when new. This is normal.

Its not as easy on some machine to get right, so expect a day or so getting it right. Others come in and are good from the get go it seems with minimal work required to get them set up. Some lathe stands are a 3 point system with 2 extra pads for taking up the slack, with the 3 main points being the main load bearing ones..

Just to avoid confusion I would like to point out that travel along a lathe's bed is usually the Z axis and travel of the cross slide is usually the X axis. This results in the Y axis being vertical. I think Neil L is looking at it like a vertical milling machine where the X and Y axes are horizontal with the Y axis being front to back but a lathe is more like a milling machine on its side.

People who use CNC on a lathe would use this usual orientation for a lathe but people who are manual users only are not likely to have picked up on this convention.

Martin C

Hi, Neil. If I can paraphrase:-

Set the headstock end of the bed via the cross slide to be level across the bed. Move to the tailstock end and tune that until it's level (across the bed). Rinse and repeat until both ends are level ...

I assume that I can set the headstock end level using the adjustment feet on the base. I will then make sure that the base is stable at the tailstock end and use shims between the base and the lathe feet to level the tailstock end.

I will give that a go.

Thanks

Iain

Iain,

Put the level on the carriage assembly, and as you wind through .from end to end you will see how it is. I recommend leaving the level on the cross slide, as that is representative of the function.

Yes I called it X and Y, But in reality, it is X is the cross slide and y is vertical to the X axis, and Z is in direction to the spindle.

Neil

Hopefully, when the Tailstock end of the bed reads the same as the Headstock end, the bed should be free of twist, and the lathe should cut parallel.

FWIW, my method of adjusting, if needed, would be to mount the lathe on short studs (with clearance in the holes in nthe lathe feet )

Each stud would be in a tapping in the steel plate, and locked into it by a nut. Above that nuty would be another on which the lathe feet sit. The lathe feet would be clamped by a further nut above each foot.

having clamped the feet at the Headstock end, thje Tailstock feet bare adjusted mto remove any twist from the bed.

Either by using the level, or using "Rollie's Dad's Method" as laid out in Ian Bradley's "The Amateur's Workshop" or his "Myford Series 7 Manual"

The lathe may not be absolutely level Headstock to Tailstock (Unless it is imperative to have coolant drain in a particular direction ), but the bed should be free of twist.

If using a level, even after checking with a rigidly mounted clock (You can't chase microns if the clock is waving about in the slightest breeze! ) I would feel inclined to check by machining, as that is the object of the exercise.

Howard

So I got out to the lathe today and tried to follow the instructions above.

My first comment is that I found it very difficult to get repeatable readings Just leaning on the bench would shift the level bubble noticeably!

Using the feet on the base I was able to get the head end more or less level (within 2 microns on the level).

I then ran along the bed and got readings like this. Note +/- infinity means off scale, + being front of the bed higher, -negative back is higher. distances are from starting point near headstock to tailstock in cm. numbers are in microns (approx).

0 : 0

5 : -∞

10 : -∞

15 : +4

20 : +∞

25 : +∞

30 : +∞

35 : ∞

To get some idea of how far out the bed is, I tried using some shims to prop up one end of the level or another. Unfortunately my smallest shim (0.05mm or 2 thou) is too large. I then tried some aluminium foil which seems to run between 15 and 20 microns in thickness. This is where I struggled with repeatability. Sometimes one thickness would reverse from +∞ to -∞ and sometimes two thicknesses didn't quite get it to zero.

Which was frustrating. I then ran down the length of the bed (at my stopping points above) to see how much foil was needed to correct the angle. Broadly I found that at all the ∞ points one or two thicknesses of foil was enough to get the level around zero or to at least reverse the direction of tilt.

Long and short of that, assuming I've not been an idiot (rash assumption that) is that a) the twist is not linear along the bed and (more important) the shim needed to correct the twist at the tailstock end is likely to be around 20 microns, but quite possibly less.

The level is 100mm and the mounting feet of the lathe bed are on 70mm centres,

Any suggestions (as always) most welcome.

Iain

How are you using this level Iain? Your readings don't make much sense. If you're placing it on the carriage and winding it up and down you are not measuring bed twist, you're measuring the relative angle of the carriage. This can be hugely affcted by wear in the bed and even more in the uderside ways of the saddle (since they will be worn much more than the actual bed ways).

If your goal is to 'level' the bed i.e. remove twist - put the level on the bed. Once you have the bedway levelled if placing the level on the carriage yields different results then you have a separate issue to correct, but at least you will be correcting it towards your baseline measurement (i.e. a bed that is properly aligned) not trying to interpret readings that are affected by several different factors.

Hi, Pete. The level is measuring across the width of the bed on top of the cross slide. the top slide has been removed so the surface is flat.

I note that Neil's recommendation is the opposite of yours! That is to use the cross slide and not the bed. That actually makes sense to me as local wear on the bed will be picked up more by the short (20mm?) width of the level vs the (comparatively) larger width of the saddle - 100mm in this case.

Surely if the bed is worn (on one edge) then I will see an angle change with the level regardless of it being on the bed or on the carriage?

I did note when I had the bed on the mill table that there appeared to be around 0.01 mm of wear along the most used part of the bed (that is nearest the headstock). By which I mean that there was a dip in this area. Which could either be wear or a twist. I suppose.

Iain

If you have or can borrow a good parallel you can check for wear in flat ways very easily. Here is a video of me doing that on my Atlas 10.

The same technique applies to all lathes with flat ways such as the Myford 7" range.

Thanks for this Benedict - good video! Sadly, my lathe has a prism on the near side .

Also, best I can tell (from sweeping the bed with an indicator when sitting on the mill bed), any errors are probably under a thou which makes feeler gauges a bit too big.

Iain

I didn't see the prism in the photos. If your wear is less than a thou I wouldn't worry about wear.

The only question then is, is there any twist. Remind me ho big is the discrepancy along a bit or turning?

I get about 2 thou over 75mm (er three inches) near the headstock. This applies to boring and finishing the outside.

Iain

Both slope inwards towards the headstock...

Iain

The problem with wear is twofold. Excessive wear of the ways will result in the tool moving away from and towards the spindle centre of rotation as the carriage goes over worn sections of the ways. It takes a lot of wear for this to show up. The more subtle problem with wear is that if the force needed to move the carriage along the ways changes it is possible the carriage can twist in the XZ plane to varying amounts as the force changes. If it is this twisting that is causing the problem then levelling the bed will probably not cure it. This twisting is more likely to show up with lathes that have a narrow carriage. My lathe is 170mm across the ways but the carriage is 280mm wide along the bed. Looking at some of the lower cost lathes they seem to have carriages with widths that are much closer to their distance across the ways.

The geometry of this twisting is that the further away from the centre of the twist the cutting point is the more it will result in non parallel cutting. The way this distance can vary with how the lathe is being used makes it hard to track it down, it's not always possible to put the cutting point close to the centre of carriage rotation.

Martin C

I just want to clarify. Is the test bar 2 thou narrower at the headstock end that at the tailstock end? I thought you said in an earlier post that it was 2 thou narrower at the tailstock end but I have just looked at your graphs and if 0 is the headstock it would appear to be narrower there. That would be what I might expect if you are turning something that long out of the headstock without tailstock support.

Have you tried it with tailstock support?

Also are you measuring the deviation by measuring the diameter of the bar with a micrometer? If so are you cleaning the carbide faces between measurements?

Thanks, Martin.

Sadly, my lathe falls in the cheaper category. The ways are about 86mm across and the carriage 100mm wide. Ther bed mounting bolts are on a 70mm centre.

I've tweaked the thing here and there and have some extra support under the ways. I recently tuned this up and have very little (under 1 thou) of lift on any part of the carriage - which moves freely until you get to the far tailstock end at which point it becomes stiff. So I have wear, but I think I've compensated for it as well as I can on a machine of this type.

Wanna swap lathes??

(no not serious!)

Iain

Headstock end, Benedict. Apologies if I mistyped earlier. I would say it's old age, but I've always been prone to typos!.

I have cut a 40mm bar in earlier tests and measured the diameter with a micrometer twice (one at 90 degrees to the other)

Yes I cleaned up the faces and as you can see I get very close agreement between the two sets of measurements. In fact most of the pairs of measurements are less then 5 microns apart.

At 40mm I would not have expected tailstock support to make much difference due to the stiffness. Also bringing the tailstock in would add an unwelcome extra factor.

Iain

If it is wider at the tailstock end, try tailstock support. Nothing is so stiff it does not bend.